| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| StrongReject | Meta-Llama-3.1-8B (ATTAQ-aligned) | ASR7 | 32 | 4d ago | |
| Harmbench | SafeKey | ASR0.3 | 32 | 4d ago | |
| SC-Safety | HS3.33 | 16 | 4d ago | ||
| Edgebench | HS Score4.53 | 16 | 4d ago | ||
| Basebench | Meta-Llama-3.1-8B (AART-aligned) | HS1.73 | 16 | 4d ago | |
| Advbench | HPR68 | 8 | 4d ago | ||
| XSTEST | HPR61 | 8 | 4d ago |